Purdue Contribution of Fusion Simulation Program Page: 34 of 204
This report is part of the collection entitled: Office of Scientific & Technical Information Technical Reports and was provided to UNT Digital Library by the UNT Libraries Government Documents Department.
Extracted Text
The following text was automatically extracted from the image on this page using optical character recognition software:
3.6.2 Criteria, policies and procedures
FSP leadership will select and prioritize computational services for production deployment based on the
following criteria:
" User demand: determination that a sufficient user base exists to warrant production deployment with
reasonable expectation of wide scale use of service;
" Readiness of software: determination that underlying scientific software is sufficiently stable and
mature to sustain wide scale use.
" Readiness of hardware: determination that appropriate platforms and data and network resources are
available and sufficiently stable to support reliable service execution at scale.
As a decision of deployment involves commitment of significant manpower and computational resources, each
decision should be based on a consensus of user groups, software development teams, and computational
resource managers. User demand is the most important criterion; in cases where user demand is not met by
readiness of software and/or computational resources, other FSP activities may be required to bridge the gap.
3.6.3 Orientation to User Support
The primary mission of the production team is to maximize user productivity by making tested, production
versions of the FSP software available to research users in as efficient and effective a manner as possible.
Research users are of course expected to have a broad understanding of their scientific applications and
research goals, but, they are not expected to have to function as code developers or code installers in order to
have access to the software. They should not (for example) be required to read Fortran or C++ source code to
determine the meanings of code inputs and outputs.
Instead, the FSP production computing team will deploy tested software on specific production platforms which
will be available to users as a service. The team will provide documented, tested methods for preparation of
input datasets with working examples. There will be a clear procedure for submission of jobs. The team will
provide tools for monitoring the progress of active simulations, and, with the help of software developers, tools
for visualization and analysis of the output of completed runs. User level documentation will be provided to
clearly explain, in detail, every step of each of these phases of run production.
The team will trouble shoot failed runs and communicate clearly with users on the status of such jobs. If cause of
failure is traceable to input data, production team members will give advice to users on input modifications. If
the cause is due to a system or hardware failure, the team will restart the job when the problem is resolved. If
the cause appears to be in an FSP component and requires further investigation by domain experts, the
production team will contact the appropriate FSP software developers.
The production computing team is oriented to serve user needs, in order to maximize the value of FSP software
for research applications.
3.6.4 Interaction with other elements in FES program
The production computing teams will interact with FSP software developers-both component developers and
the general software infrastructure and support team. Production computing necessarily has a very strong
concern for software quality, and FSP is investing significant resources to insure quality. Nevertheless, it is
inevitable that production deployment will reveal issues with FSP components that will require attention of
component experts going beyond what is available within the teams themselves. In such situations, operations
teams will rely on FSP software developers with appropriate component expertise. The role of the operations
teams will be to provide the necessary data and information to enable FSP component specialists to precisely
reproduce problematic code behaviors, as will be required to enable isolation and repair of trouble spots. Where
this is not possible, it will likely be related to software quality issues or conformance to standards, and, the
production team will likely need the help of the software infrastructure and support organization.
The production teams will rely on the software integration and support team for robust and reliable build
procedures - as will be required to reliably roll out new versions of production software. The production team27
Upcoming Pages
Here’s what’s next.
Search Inside
This report can be searched. Note: Results may vary based on the legibility of text within the document.
Tools / Downloads
Get a copy of this page or view the extracted text.
Citing and Sharing
Basic information for referencing this web page. We also provide extended guidance on usage rights, references, copying or embedding.
Reference the current page of this Report.
Brooks, Jeffrey. Purdue Contribution of Fusion Simulation Program, report, September 30, 2011; United States. (https://digital.library.unt.edu/ark:/67531/metadc844493/m1/34/: accessed April 25, 2024), University of North Texas Libraries, UNT Digital Library, https://digital.library.unt.edu; crediting UNT Libraries Government Documents Department.